Joule Ready Case Study

I spent 2.5 years working with Brian McCracken to find a way to work consumables into the ChefSteps business. Together we ran several experiments to learn what a food product for our community would look like, essentially running a startup within a startup. I helped build and run the Joule Ready team for ChefSteps—Our product was still in early stages and was growing at a good rate with high customer retention when ChefSteps laid off 90% of staff, including our entire Joule Ready Team.

Joule Ready

How did Joule Ready start?

Before ChefSteps even set out to compete in the sous vide device market, in 2014 we started thinking about providing meal kits to people for ChefSteps recipes. We put together a kit for the sous vide pastrami recipe that is highly technical, labor intensive, and downright delicious. We knew we had a highly dedicated audience when we sold these kits to buyers in a short amount of time at a high price point. We also knew that the margin, cost of goods sold, and labor costs would be too high to justify the product. We shelved it for a couple of years until after we established Joule as a sous vide competitor, knowing that someday we would come back to it.

In 2016 we shipped Joule and a year later we started working on exploring options for a consumer packaged good to complement Joule Sous Vide. We ran a plethora of experiments to try and find out the right combination of price, convenience, and quality. We systematically went through different combinations of variables in order to learn what was most important to our customers.

Butcher Shop Pop-ups — We worked with B&E Meats to create packages of sous vide-ready meats that were ready to cook with Joule. We worked with both pickups and deliveries to customers, and the bundles cost up to $179 for a wide variety of meats. Main variables tested: Delivery Logistics (Convenience), Pricing Models (Bulk buying)

Roast Sale with Heritage Meats — One of the most telling experiments we had was our roast sale that we held at ChefSteps before the big eating holidays in the States. We sold ~100 roasts and large cuts of meat, many of which were "upgraded" to sous vide-ready roasts for $5. An overwhelming percentage ( ~75%) of our cuts were upgraded with garlic, herbs, salt and pepper. The margin was there, the demand was there. This was the turning point that led us to believe that creating sous vide-ready meats was our main product. The next year or so would be learning how to create that product. Main variables tested: Pickup Logistics and sous vide-ready Packaging (Convenience)

Pick-ups with Heritage Meats — We ran pickup points in Seattle and Bellevue for customers who purchased meat. The pickups were largely based off of what Zaycon Fresh was doing at the time, but specificaly for higher-end meats. We found that while pickups can be convenient, the pricing structure wasn't conducive to feeling that way—High-end meats in bulk are cheaper, but our customers wanted to pay for convenience and quality. Picking up meat just wasn't in the cards for these customers. Main variables tested: Pickup Logistics (Convenience), Quality/Price (High-end cuts bought in bulk)

Sous Vide Brisket Kit with Snake River Farms — Snake River Farms was interested in working with us, and we had an amazing recipe for brisket sitting on our website. We knew the best way to get people cooking was to provide ingredients along with the skills, time, and tools required to make a delicious piece of meat. We created a brisket kit out of the recipe and our connections with Snake River Farms. We tested out dropshipping with SRF and included a bunch of stuff along with it. We sold out within days. Main variables tested: Delivery Logistics (Convenience)

Smokerless Smoked Brisket Kit

On-Demand BBQ — We made a large batch of sous vide-cooked meats and BBQ, put them in the deep freeze and sold it as On-Demand BBQ, fulfilled by Postmates. The order sizes were small and the delivery fee was large—We were optimized for convenience first and foremost in this experiment. Main variables tested: Delivery Logistics and Ready-to-eat Meats (Convenience)

Brian and I worked together to conduct most of these experiments with help from developers when needed. He made business connections and was ground game, pounding pavement to develop relationships with meat purveyors, butchers, and processors. I did most of the experiment setup, user interviews, UX and product design, and learning roadmap to help us refine strategy. We worked with Jeremy, a developer to help us use Shopify to sell stuff. I designed the product pages and galleries for the e-commerce flows as well as product labels.

After a year of R&D, we finally settled on a product:

Joule Ready (Sauces in Bags) — Meat delivery was hard. Logistics of frozen or refrigerated foods made shipping very expensive and made our margins shrink. People wanted the meat, but we wondered if they were willing to pay a premium for just that sous vide-ready bag with sauce already inside. We started with herbs, spices, garlic, etc. and quickly moved to package dry rubs, then finally to liquid sauces. One of the first things we tried was just packaging our BBQ spice rub recipe into the bag with pork shoulder and have that come out as an easy 24-hour pulled pork—The active prep time was less than a minute. The sauce-in-a-bag idea was Brian's and I thought was pretty brilliant; we ran with it.

Why didn’t that other stuff work?

By the time we had settled on our business model of sauces in bags, we had already seen the influx of meal-kit startups everywhere (by the time we were done, most of them had gone as well). We knew from earlier experiments that margin was weak in providing full-meal solutions. We were determined not to inherit the cost structure of these types of places—of providing tiny packages, shipping, and extensive ingredient sourcing—but also the growth constraints from the cooking fatigue people faced from having to choose from a similar set of menus every week. For our particular audience, cooking creatively was a core tenet so we had to have that integrated into the experience.

What happened in the first year of JR?

During the first 6 months of Joule Ready, we ran an extensive beta to 120 of our core community. We used these folks to test multiple parts of our product: Packaging, sauce flavors, delivery mechanism, app and shopping UX, marketing and positioning. The only thing we couldn't really test realistically was the biggest: Price. We worked hard on offering the best product, but decided we needed to ship at a real price point to really understand if our product was going to sell.

In September of 2018, six months after beta, we soft-launched the product under the name of Joule Ready Preview. In 3 short months, Brian had built out a facility for piloting our sauce line and shipping them. We had loads of data, user feedback, product knowledge, and a strong stable of sauces to go to launch with. We started with 12 different sauces and quickly gained traction, launching to a limited group of users. After a month of preview testing, we made the sauce generally available.

Initial growth was awesome, 50% month-over-month for the first 4 months, all through email as our only sales channel. We hadn't yet developed our social voice, done any PR, or completely worked out our content marketing strategy.

During the next 3 months, we shifted our focus to growth.

What was the growth strategy for JR?

It wasn't easy at first—Our team was scattered and needed clearer direction and wanted more strategy to rely on when creating product. I understood this and helped craft a framework for us to focus on creating good product and understanding how we were going to reach more customers. We continued to do research (with Kristina Voros helping lead the charge) about why customers were and were not buying our products. I'd been here before with the early days of ChefSteps and wanted a clear path forward as well, so went back to using Pirate Metrics in order to elucidate.

I worked on communicating our KPIs and setting realistic goals for our team.

Our first- and second-quarter learning roadmaps were primarily focused on the upper part of the funnel. We lagged in acquisition and were mining our email lists. We experimented with a few different types of content that would move the needle, changing the type, frequency, and depth of each piece. We arrived at a good framework for content, but were always looking to learn more. The types of content were:

What could we have done differently?

Looking back, there are a lot of moments of opportunity. We had clearly defined problems for the user and crafted our product as a good solution for those—We had placed a lot of emphasis on the Jobs to Be Done for the product. But our integration with marketing came late in the process, so our acquisition and growth strategy were lagging; we tried to do too much in too little time, and our communication process to customers wasn't clear enough to allow them the see how our product fit into their lives from the beginning. We over-indexed on convenience and didn't focus on our community's one true desire: To cook like chefs. One of the flaws in soft-launching was that we weren't able to really keep momentum much longer than a few months, and only from the same set of users in our email.

What were the biggest problems?

We conducted more research into why our market penetration amongst Joule users was lower than we would have liked, and the answers came out obvious and telling. The main two concerns people had were price and their willingness to use the product—Most non-buyers insisted that they would rather cook from scratch.

This essentially bifurcated the product use: People that were willing to pay the price were also people that didn't mind using (or admitting to using) a "shortcut" in their cooking. Many customers that did not want to pay the price were looking for convenience that they could get much more cheaply—This placed us firmly in competition with jarred, ready-to-eat sauces. People that wanted to make their own sauce were cooks that wanted to cook, and weren't looking for shortcuts. We had a clear problem of communicating this to users and much of our future content plans were made to alleviate that concern.

What was the JTBD for JR?

Initially we thought that the job Joule Ready would be useful for would be a convenient weeknight dinner for our customers. Joule was already being used on the weekend for project cooks, so we were attempting to capture one more night a week in order to increase overall engagement with Joule Sous Vide. When customers want to use Joule to cook an easy, convenient dinner, they will use Joule Ready because it's convenient and still delicious—They will still get the convenience without compromising quality.

This, however, didn't turn out to be the case. The cooking engagement that we saw still happened on the weekends—Usage would spike every Sunday night. Instead of people cooking more, they were cooking the same amount, except they were using Joule Ready to cook. While this isn't the behavior we were looking for, we knew that would increase the customer lifetime value if we could count on that customer to continue to buy and use Joule Ready multiple times over the year.

Lessons Learned

I still contend that shipping fast was the best thing we did here. Getting to a place where we were capturing real data on purchases and getting customer feedback quickly was essential to the success of the product. We simply did not have enough time to refine our marketing and acquisition strategy, but if we had done that from the beginning, we would have learned in that area just as quickly as we did about the product and customers. Our marketing and sales were just getting into gear and I know if we had more time we could have solved that problem as well.

  • Product Team Work